>REM SUBMIT 'D:\mydocs\ys209\incxed.syc' /ECHO
>mglh
>select sex$='Male'
>let hsdro=0
>if educatn=1 or educatn=2 then let hsdro=1
>let hsgra=0
>if educatn=3 then let hsgra=1
>let somco=0
>if educatn=4 then let somco=1
>let cogra=0
>if educatn=5 then let cogra=1
>let grade=0
>if educatn=6 or educatn=7 then let grade=1

>model income=constant+educatn
 

MGLH - Statistics

>estimate

Data for the following results were selected according to:

      sex$='Male'

Dep Var: INCOME   N: 104   Multiple R: 0.473796   Squared multiple R: 0.224482

Adjusted squared multiple R: 0.216879   Standard error of estimate: 14.528751

Effect         Coefficient    Std Error     Std Coef Tolerance     t   P(2 Tail)
CONSTANT          4.879179     3.962622     0.000000   .        1.23130  0.22104
EDUCATN           5.371633     0.988578     0.473796  1.000000  5.43370  0.00000

                             Analysis of Variance

Source             Sum-of-Squares   df  Mean-Square     F-ratio       P
Regression           6232.282646     1  6232.282646   29.525044    0.000000
Residual             2.15306E+04   102   211.084616

-------------------------------------------------------------------------------
Durbin-Watson D Statistic     1.968
First Order Autocorrelation   0.012
>

Plot of Residuals against Predicted Values


>model income=constant+hsgra+somco+cogra+grade
 

MGLH - Statistics

>estimate

Data for the following results were selected according to:

      sex$='Male'

Dep Var: INCOME   N: 104   Multiple R: 0.476472   Squared multiple R: 0.227025
Adjusted squared multiple R: 0.195794   Standard error of estimate: 14.723041

Effect         Coefficient    Std Error     Std Coef Tolerance     t   P(2 Tail)
CONSTANT         13.388889     3.470254     0.000000   .        3.85819  0.00020
HSGRA             7.841880     4.195332     0.232359  0.505263  1.86919  0.06455
SOMCO            15.905229     4.979329     0.359977  0.614778  3.19425  0.00188
COGRA            17.548611     5.058721     0.387521  0.625668  3.46898  0.00078
GRADE            24.825397     5.246531     0.518600  0.650000  4.73177  0.00001

                             Analysis of Variance

Source             Sum-of-Squares   df  Mean-Square     F-ratio       P
Regression           6302.888552     4  1575.722138    7.269166    0.000036
Residual             2.14600E+04    99   216.767928

-------------------------------------------------------------------------------
Durbin-Watson D Statistic     1.933
First Order Autocorrelation   0.028
 

Plot of Residuals against Predicted Values


>rem one can also omit the constant and include all the indicators
>model income=hsdro+hsgra+somco+cogra+grade
 

MGLH - Statistics

>estimate

Data for the following results were selected according to:

      sex$='Male'

Model contains no constant
Note the inflated R-Square when constant is not included.  This is called non-centered R-Square and should
not be used as a measure of fit.

Dep Var: INCOME   N: 104   Multiple R: 0.876517   Squared multiple R: 0.768283
Adjusted squared multiple R: 0.758920   Standard error of estimate: 14.723041

Effect         Coefficient    Std Error     Std Coef Tolerance     t   P(2 Tail)
HSDRO            13.388889     3.470254     0.186657  1.000000  3.85819  0.00020
HSGRA            21.230769     2.357573     0.435674  1.000000  9.00535  0.00000
SOMCO            29.294118     3.570862     0.396889  1.000000  8.20365  0.00000
COGRA            30.937500     3.680760     0.406639  1.000000  8.40519  0.00000
GRADE            38.214286     3.934898     0.469844  1.000000  9.71163  0.00000

                             Analysis of Variance

Source             Sum-of-Squares   df  Mean-Square     F-ratio       P
Regression           7.11530E+04     5  1.42306E+04   65.648987    0.000000
Residual             2.14600E+04    99   216.767928
-------------------------------------------------------------------------------
Durbin-Watson D Statistic     1.933
First Order Autocorrelation   0.028
 

Plot of Residuals against Predicted Values



>rem now same comparison with logged income as dependent var
>let l10inc=l10(income+1)
>model l10inc=constant+educatn
 
 

MGLH - Statistics

>estimate

Data for the following results were selected according to:

      sex$='Male'

Dep Var: L10INC   N: 104   Multiple R: 0.462675   Squared multiple R: 0.214068
Adjusted squared multiple R: 0.206362   Standard error of estimate: 0.282933

Effect         Coefficient    Std Error     Std Coef Tolerance     t   P(2 Tail)
CONSTANT          0.936115     0.077168     0.000000   .       12.13085  0.00000
EDUCATN           0.101473     0.019252     0.462675  1.000000  5.27088  0.00000

                             Analysis of Variance

Source             Sum-of-Squares   df  Mean-Square     F-ratio       P
Regression              2.223991     1     2.223991   27.782171    0.000001
Residual                8.165203   102     0.080051

-------------------------------------------------------------------------------
*** WARNING ***
Case          237 is an outlier        (Studentized Residual =    -3.646594)

Durbin-Watson D Statistic     1.805
First Order Autocorrelation   0.093
 

Plot of Residuals against Predicted Values



>model l10inc=constant+hsgra+somco+cogra+grade
 

MGLH - Statistics

>estimate

Data for the following results were selected according to:

      sex$='Male'

Dep Var: L10INC   N: 104   Multiple R: 0.492572   Squared multiple R: 0.242627

Adjusted squared multiple R: 0.212026   Standard error of estimate: 0.281922

Effect         Coefficient    Std Error     Std Coef Tolerance     t   P(2 Tail)
CONSTANT          1.033201     0.066450     0.000000   .       15.54866  0.00000
HSGRA             0.248897     0.080334     0.381243  0.505263  3.09830  0.00253
SOMCO             0.376642     0.095346     0.440662  0.614778  3.95028  0.00015
COGRA             0.381972     0.096866     0.436039  0.625668  3.94330  0.00015
GRADE             0.511031     0.100462     0.551855  0.650000  5.08679  0.00000

                             Analysis of Variance

Source             Sum-of-Squares   df  Mean-Square     F-ratio       P
Regression              2.520699     4     0.630175    7.928745    0.000014
Residual                7.868495    99     0.079480
-------------------------------------------------------------------------------
*** WARNING ***
Case          237 is an outlier        (Studentized Residual =    -3.643266)

Durbin-Watson D Statistic     1.764
First Order Autocorrelation   0.113
 

Plot of Residuals against Predicted Values


>REM End of command batch file D:\MYDOCS\YS209\INCXED.SYC
 
 


Last modified 14 Mar 2000